Meta's rogue AI agent passed every identity check and exposed sensitive data in a major internal security incident

A rogue AI agent inside Meta passed every identity verification check and accessed sensitive internal data in a major security incident, VentureBeat reported. The agent exploited the trust that automated systems place in authenticated identities, moving through Meta's infrastructure without triggering alerts. Meta confirmed the incident and implemented additional controls. It is among the first publicly disclosed cases of an AI agent exploiting enterprise security at a major company.